And still they gazed, and still the wonder grew, that one small head could carry all he knew. Oliver Goldsmith

A tradition without intelligence is not worth having. T. S. Eliot

Debating ideas in education – and anywhere else – is essential if we want to improve the lot of children and society. Over the past 6 years of so I’ve learned huge amounts from taking part in this back and forth and have, as well as becoming a good deal more knowledgeable, become a lot more adept at thinking critically about the ideas I encounter. My views have changed a great deal over this period and so now seems like a good time to lay out a summary of my hypotheses about education.

  1. Everyone seems to agree that increasing children’s creativity, problem solving, critical thinking etc. is a worthy goal of education. Like everyone else, I want to see children able to make their way in an uncertain world, but I disagree with the consensus view about the best way to achieve this aim.
  2. Intelligence is a social good (a greater number of individuals with higher IQ makes society safer) as well as an individual good (IQ correlates strongly with creativity, leadership, happiness, longevity etc.). It therefore seems reasonable that if we want to make children more creative and better critical thinkers, we need first to make them cleverer. See here for my views on IQ and intelligence.
  3. Although many of the claims attached to growth mindset are wrong (the brain is not like a muscle) we can increase IQ. The Flynn effect shows how aspects of IQ tests respond to cultural change and is rising by about 3 points per decade.
  4. Intelligence can be broken into fluid intelligence and crystallised intelligence. Fluid intelligence is our raw reasoning power, and is, as far as we can tell, fixed; nothing we’ve tried as yet is able increase it. Crystallised intelligence is the ability to apply what you know to new problems and can certainly be increased by adding to our store of knowledge.
  5. No one can think about something they don’t know. Equally, the more you know about a subject, the richer and more sophisticated your thinking on that subject becomes. It is my view that the much vaunted ’21st century skills’ depend on knowing stuff within a specific domain.
  6. Evolutionary psychology points to the fact that some things are easy for us to learn (biologically primary evolutionary adaptations) but some are hard (biologically secondary modules). So-called ’21st century skills’ are actually biologically primary adaptations and cannot be taught as generic skills. They are in fact domain-specific and rely on expertise in a domain in order to be exercised and developed.
  7. Fluid intelligence is strongly correlated with working memory capacity (which also appears fixed) but its limits can be ‘hacked’ by storing information in long-term memory. Fortunately, the capacity to store new information does not appear to be correlated with our fluid intelligence: all children can remember stuff, regardless of how able we perceive them to be.
  8. When we store knowledge in long-term memory, it organises itself into schemas which, when we use them to think about complex problems, take up less of our limited working memory capacity.
  9. Declarative knowledge (facts) is what we think about; non-declarative knowledge (skills) are what we think with. Certain kinds of procedural knowledge (skills) can be automatised so that they takes up practically no space in working memory, leaving us with far more capacity with which to think. I tried to show the relationship between memory about skill in my taxonomy.
  10. Certain domains of procedural knowledge are particularly worth automatising because they recur so often in education – phoneme/grapheme relationships, times tables facts etc. The process of automatisation is accelerated through purposeful practice.
  11. It may be true that “everything works somewhere and nothing works everywhere,” but if so it’s trivially true. Better to say, some things work in most contexts and other things rarely work anywhere. ‘Traditional’ approaches to the curriculum and instruction are better suited to achieving the ends most people value than ‘progressive’ ones. This is, I think, a social justice argument.
  12. If children automatise powerful procedural knowledge in long-term memory and encounter culturally rich declarative knowledge, they will become cleverer and, therefore, more creative, better problem solvers etc.
  13. Because intelligence (and therefore everything else) relies to some degree on genetic heritability, we cannot hope to close the gap between the most and least able. We can, however, move the whole bell curve to the right which would both benefit society as a whole and every individual within it.
  14. We may also be able to change the shape of the curve. If we try to develop skills by trying to teach generic skills directly, then children with higher fluid intelligence and those from more advantaged backgrounds will be privileged and the bell curve will stretch out. But, if we make a concerted attempt to increase children’s crystallised intelligence then the curve might become steeper as the difference in what children knew became less great. (This is the bit I’m least sure about – critique welcomed.)

I would be genuinely grateful for any constructive criticism of any of these steps in my chain of reasoning, or if anyone can suggest vital bits of information I might have overlooked.